PC@LIVE wrote on 2025-01-28, 15:22:

I did a boot test, first with the Pentium 133, later I changed CPU, and I put a Pentium 200 MMX, with this I made measurements, […]
Show full quote

I did a boot test, first with the Pentium 133, later I changed CPU, and I put a Pentium 200 MMX, with this I made measurements, to look for VCORE lower than 2.8V, the PC always started, but the measurements on the Mosfet or in the inductor, were all quite strange, I have to try them again, because even changing position to the jumpers I see a strange 3.06V, even when it should only do 2.2V, I think I have to find the PINs in the back of the CPU (socket 7), and see what's really there, usually that type of measurement works very well, but Here maybe there are other components in the VCORE line, which could lower the voltage of 3.06V to the selected one.

This card does not have the VCORE self-detection function, and there is no BIOS function that can modify it, now I do not remember if there is a hardware monitor in the BIOS, with the relative voltages, I will see if there is and what it shows.

The Intel Pentium MMX 200 CPU, is of the blocked type, it only makes 166 and 200 MHz, so it is not possible to overclock at 233 and underclock at 133.

For now since the VCORE is high, I will not try any AMD K6-2 CXT, only when I am sure that the 2.2V corresponds to something similar, I will be able to do the CPU change, and redo the tests with undocumented VCORE settings.

The BIOS version is the factory one (?), it reads BT-1B2, and the motherboard should have a 1 Mbit BIOS chip, I will see later to update it to the BT-1B7 version, if I can make some progress with the undocumented VCORE.

Hi

I have this Soyo motherboard SY-5BT and it is currently with a pentium 133 but it currently displays on the POST screen BT-1B7 and the bios date is 02/04/1999
The full POST string is: 02/04/1999-i430tx-w977-2A59IS2BC-00

The setup program is full of tweaking options and this version doesn't seem to be on the retro web motherboard site

Is this the correct BIOS for this board? some "hacked" unofficial version ?

Hi Nunoalex,

This BT-1B7 BIOS from 02/04/1999 is indeed the latest official release from Soyo for the SY-5BT Ver 1.x board, and is available on TRW at https://theretroweb.com/motherboards/s/soyo-s … 5-ver.-1.1#bios

The latest, and in my opinion best, 5BT BIOS is the 1Mbit 03/15/2000 Rel 2.2 BIOS upgrade by Unicore Software. It is on TheRetroWeb too.
This BIOS has full IDT, Intel, Cyrix, and AMD CPU support including the K6-2+/III+. It has several other improvements like boot from CDROM and support for HDDs up to 128GiB. And of course it has a fix for the Windows 98 UDMA bug. 😉

Cheers, Jan

In the last few days I've dedicated my time to something else, however in the next few it will be like that, I can't always find a few minutes to dedicate to oldware.

Today, however, there is a novelty, I received an assortment of 1/4W Resistors, in total there are 2600 PCs (130 different values), ranging from 1 OHm 🕉 to 3 MOHm, what is this assortment for?

Well, in the meantime I will be able to avoid searches on scrap boards, moreover you can almost never find the value you are looking for, then I will finally be able to finalize, the experimentation on the Soyo SY-5BT, for VCORE lower than 2.2V, currently I have found the values of the resistors, to select from 2.3V to 2.7V, for those lower from 2.1V down, I have to pass the resistors of the VCORE jumpers, and try various values, all below 39K, what interests me is to get around 1.6V, but this depends on the minimum voltage that the motherboard can do, some do not go below 2.0V others yes They stop at 1.8V, and others even reach 1.3V, I don't know how far it's possible to go down, I wouldn't mind getting to 1.5-1.6V, I wouldn't mind at all if it stops at 1.8V, I'd be quite satisfied if I can reach 1.3V, what's the point of all this?

I want to solve the problem of the Cache L2 chip, with an AMD K6-2+ 400 1.6V, the integrated cache of 128KB (which would even be possible to increase to 256KB), replaces that of the motherboard that becomes L3, and above all bypasses many of the limits of the cache soldered on the motherboard, in short the only thing is that it would work at FSB 66 instead of 100, but if you want something more you can switch to FSB 75, speed that strangely Intel never supported.

In addition to the various resistors, I received a transparent box with closure, and a very useful summary table of color codes, which can be easily consulted, in case of doubt, however with this stock I think I should no longer buy 1/4W resistors, and according to what I see they are also quite precise, compared to the most common ones with a 5% tolerance (here instead 1%).

Nunoalex wrote on 2025-02-09, 22:29:

Hi […]
Show full quote

PC@LIVE wrote on 2025-01-28, 15:22:

I did a boot test, first with the Pentium 133, later I changed CPU, and I put a Pentium 200 MMX, with this I made measurements, […]
Show full quote

I did a boot test, first with the Pentium 133, later I changed CPU, and I put a Pentium 200 MMX, with this I made measurements, to look for VCORE lower than 2.8V, the PC always started, but the measurements on the Mosfet or in the inductor, were all quite strange, I have to try them again, because even changing position to the jumpers I see a strange 3.06V, even when it should only do 2.2V, I think I have to find the PINs in the back of the CPU (socket 7), and see what's really there, usually that type of measurement works very well, but Here maybe there are other components in the VCORE line, which could lower the voltage of 3.06V to the selected one.

This card does not have the VCORE self-detection function, and there is no BIOS function that can modify it, now I do not remember if there is a hardware monitor in the BIOS, with the relative voltages, I will see if there is and what it shows.

The Intel Pentium MMX 200 CPU, is of the blocked type, it only makes 166 and 200 MHz, so it is not possible to overclock at 233 and underclock at 133.

For now since the VCORE is high, I will not try any AMD K6-2 CXT, only when I am sure that the 2.2V corresponds to something similar, I will be able to do the CPU change, and redo the tests with undocumented VCORE settings.

The BIOS version is the factory one (?), it reads BT-1B2, and the motherboard should have a 1 Mbit BIOS chip, I will see later to update it to the BT-1B7 version, if I can make some progress with the undocumented VCORE.

Hi

I have this Soyo motherboard SY-5BT and it is currently with a pentium 133 but it currently displays on the POST screen BT-1B7 and the bios date is 02/04/1999
The full POST string is: 02/04/1999-i430tx-w977-2A59IS2BC-00

The setup program is full of tweaking options and this version doesn't seem to be on the retro web motherboard site

Is this the correct BIOS for this board? some "hacked" unofficial version ?

For your question, Jan answered you very well, he also suggested to me to use the BT-1B7 version, which seems to be the one with the best CPU support, and large capacity HD.

Chkcpu wrote on 2025-02-10, 14:39:

Hi Nunoalex, […]
Show full quote

Hi Nunoalex,

This BT-1B7 BIOS from 02/04/1999 is indeed the latest official release from Soyo for the SY-5BT Ver 1.x board, and is available on TRW at https://theretroweb.com/motherboards/s/soyo-s … 5-ver.-1.1#bios

The latest, and in my opinion best, 5BT BIOS is the 1Mbit 03/15/2000 Rel 2.2 BIOS upgrade by Unicore Software. It is on TheRetroWeb too.
This BIOS has full IDT, Intel, Cyrix, and AMD CPU support including the K6-2+/III+. It has several other improvements like boot from CDROM and support for HDDs up to 128GiB. And of course it has a fix for the Windows 98 UDMA bug. 😉

Cheers, Jan

Hi Jan, I am very happy 😀 that you are here, on the page where I write about various MB, currently if you have read the previous messages, I have successfully experimented on the Soyo 5BT5 the VCORE from 2.3 to 2.7V, in the next few days I should be able to go below 2.2V, I hope to be able to get even below 1.6V, which was the starting goal, I know it works well with VCORE 2.0V, but I should be able to find the resistors for lower voltages, when I have completed the experimentation, I will start the construction 🏗 of an additional board for the selection of the desired VCORE, any suggestion is very welcome, the results will be within the reach of those who have a minimum of welding skill, unfortunately 😣 in this tab adding a resistor to the jumper, the voltage rises 😬.

There is one thing I never understood, that is what it is for, on the manual of the Soyo SY-5BT there is no written, it is about the JP10 jumper, looking better at this jumper and the tracks, it seems that one goes towards a Mosfet (?), at this point I wonder what effect it could have, if you put a jumper on JP10, I don't remember if I had done a test in the past, however I will try to understand if it connects to something, like if one of the two pins goes to ground, looking closely it seems to be able to affect the VI/O, so I would rule out that it has any effect on the VCORE, maybe it could be a jumper to lower the voltage on the VI/O?

Unfortunately, not being very competent in electronic circuits, I can't get ideas about what it can do, or what effect closing the two PINs can have, certainly it could be easy for those who have some competence in the field.

Maybe it's a circuit to allow the installation of Tillamok, because at least in Rev.2. X there is the 2.2V jumper, and that JP10 could for example lower the VI/O voltage below 3.0V, if that were the case it would be very interesting, but there is a possibility that it will instead serve a K6-300 CPU, there is a version with VCORE 2.2V and VI/O 3.45V, maybe JP10 can raise the voltage from 3.3V to 3.45V, maybe who knows?

Today, I can work on an AZZA motherboard (which mainboard!, said an advertisement of the period), so going back to the topic, I have this AZZA P4M2-MVB REV:1.1 between my paws, uATX format with integrated video and audio, but no one prevents the use of special cards as an alternative, since there is an AGP slot and three PCI, one of the PCI could be used to put a LAN card (there is no integrated), and even for a modem-fax, in case you want the PC to communicate with the Internet, or simply use the telephone line to send or receive faxes.

However, if the PCI slot needs free, we will have to get an external modem-fax, USB or even one of those old models that connected to serial ports (at least I hope I don't get the wrong port).

This card has a VIA chipset and two DDR RAM banks, the BIOS should be a Phoenix, so I read in the label 🏷, the integrated sound chip is an ALC201A, for the ATX power supply there are three molex, one 20 PIN one 4 PIN and one for 5.25” or HD CD drives, I think it is not possible to use an ATX power supply with 24 PIN, because on one side there is an electrolytic capacitor attached, if the additional pins are on that side (?), a 24-20 PIN adapter should be used.

The problem to be solved is the presence of numerous swollen electrolytic capacitors, a total of thirteen from 1000uF 6.3V, plus another from 3300uF 6.3V in the CPU area, of those there are seven in total, of which six in a row, the swollen one is between the six, I decide to remove all six, and leave the one attached to the ATX 20 PIN connector.

After the removal, I checked each capacitor, those of 1000uF 6.3V, are all at a loss and with a high ESR, four even have a very high ESR (up to 13.4 OHm 🕉), this makes you think that the board cannot start, or if it did, it had major stability problems, as replacements I have Chinese capacitors, Chongx brand, I don't usually use them on the MBs, but the quality ones, I don't have enough available, meanwhile these should allow me to make the board work, then in the future I could change them, but the board will be used very little, Just a few benches, or some tests, so much so that I was thinking of putting on a Celeron 478, since in sockets 478 I only have Pentium 4.

Those from 3300uF 6.3V, will be replaced with Nichicon HM, the removed ones were slightly above the tolerance (capacity), one was at a loss (swollen), and another was open, this is possible that it was damaged during the removal, because one of the poles is stretched, I probably pulled a little too much.

The last job I did is to free the holes from the tin, well for those from 1000 no problem, for those from 3300 instead the negative can't be released, it's something that often happens in cards like Gigabyte or some ASUS and ASROCK, the tin melts, but as soon as you try to suck it up, it cools down and the hole remains clogged, maybe it should be heated before freeing it? If there are any suggestions, I listen to them gladly, and I appreciate them very much, because I find it strange that there is all this difficulty, sometimes even in re-welding I have found problems, with the tin that does not stick to the MB.

I'm at the end of the work on the MB AZZA, the seventh capacitor from 3300uF 6.3V I don't change it, I think if it's like the others, it should be slightly above the tolerable capacity, honestly I think it's not very important, if it was a slightly higher ESR, it could have been a problem, however the replacement of the other six, should make those circuits work again, and for me it's the most important thing, the biggest problem was those of 1000uF 6.3V, all will be replaced by the Chongx, I'll try some with my ESR70, just to avoid putting some Defective, on average they have an ESR of 0.20 OHm 🕉, I think it's a normal value for similar electrolytic capacitors.

After the welding work, that is, replacing all the electrolytic capacitors with the new ones, I should look at some traces under the microscope, they have I don't know how deep marks, and they are near the unlocking stick, maybe they are the result of a forcing with the screwdriver, during the removal of the coupling for the CPU heatsink (plus fan), I hope they are fine, and that they are not interrupted.

The card completed the checks, would be ready to be tested on the bench, having everything indispensable (integrated), it only remains to add a RAM and a CPU and its heatsink plus fan, then just connect the ATX power supply and press PWR-ON, usually you can cheer if we see post codes scroll, and the BIOS screen appears on the video, if instead on the POST tab we see — —, words containing various animals often come out. 😉

I forgot, after a patient work, I freed the holes of the 3300 electrolytic capacitors, they are all on the side in correspondence with the + of the capacitor, and so now I can proceed with the welding of the new ones, when I finish I will show you the final result, I don't know if I will have time to try it, probably not, in the meantime I will look for a CPU that can be good, I thought of a Celeron of over 2.5 GHz, I remember having a 2.6 GHz but I don't know if I have something faster.?

Update on the MB AZZA P4M2-MVB, I finished the replacement and welding work a little while ago, and so the board should be ready to be tested, but you have to look at the microscope at the two marks on the tracks (near the unlocking lever), maybe they are not so damaged, but by eye 👁 it is not possible to understand it, by enlarging you will understand if the mark is superficial or deep.?

The chipset as I had anticipated, is a VIA P4M266, I think it is the first chipset for P4 with DDR support produced by VIA, I don't expect better performance than Intel analogues, barring surprises

The idea for this MB, is to make a PC with Celeron 478, not because I miss the P4s, but because I essentially don't have any PC 478 Celeron, and this in my opinion could be the ideal card, for DDR and CPU FSB 400 support, a Celeron on MB FSB 800 in my opinion would be a waste (even of potential overclocks).

Probably the board will be working again, there is only one SACON capacitor from 3300uF 6.3V, plus three others from 1500uF 16V, I don't think it's necessary to replace them for the moment, in the future maybe yes, however I've already replaced all those KO, plus some of those near the CPU, and for the moment it's fine, however I noticed that there are a couple of empty jumpers, JP1 JP3 and JP4, by default there should be a jumper on 1-2, currently there is no jumper, I'll fix it by putting them.

A motherboard on which I would like to finish the work, is this ATC-1411B with UMC U5SX 486-33F CPU, the problems it has are quite visible, there is green oxide in many places, mostly near the battery, but also on the ISA slots and on the sockets of the TAG and Cache chips, it is not possible to use the RAM 72 PIN, because on the one hand the hook is broken.

But looking carefully, something I had already done in the past, I noticed that two SMDs are missing, one seems to be initialed C18, and the other is R71, unfortunately I don't have other similar cards to understand the value of R71 (reading it), for C18 instead I think it's like those near C16 and C17, I don't know the values, nor how to measure them, if I'm not mistaken my ESR70 can't read them, at this point, I could do some further cleaning work, even if some points are corroded, and they should be revived and tinned.

The 72 PIN RAM socket, instead, should be replaced with a new or recovery one, fortunately there are four 30 PIN RAM sockets that seem to be in place, so that work should be done only after bringing the card back to work.

In that case, you have to find the TAG chips and cache, the TAG chip is easier to find, because in many motherboards it is still ️ used, I'm talking about cards like 370 Slot1, the L2 cache was brought from the MB to the CPU, strangely the TAG chip was not!, then of course you have to see if you can install it on this 486, theoretically it should be a little faster, apart from this in this card the TAG chip is a 62256 and has 28 pins, while those caches are 611024 with 32 pins, but remember that in the L2 cache sometimes you They could use chips with 28 pins, in fact if I'm not mistaken some recommended the installation.

I go back to the experimentation of the Soyo SY-5BT, making a brief summary, I can select all the voltages from 2.2V to 2.9V plus that of 3.2V, if I want I could try those of 3.0V and 3.1V, but at the moment I am not interested in these voltages, which I will try later, because they could be useful for any overclocking.

If I have the necessary time, this weekend I will try to lower the VCORE, and I will go down to 2.1V and gradually 0.1V to where it is possible, if it goes down I will try voltages that I do not need, and they would be those from 1.5V to 1.3V, they could be useful to try low clock speeds and to make a PC with a fanless heatsink, or possibly with a slow fan (very silent), in short, the ideal if you want to underclock.

After this experimental work, I should start the construction 🏗 of the additional card, for the selection of the desired VCORE, from there I will be able to select all the voltages from 1.3V (or more depends on the minimum one), up to 3.1V, in short it will not be anything complicated, just move the jumper to get the VCORE of the CPU, or in case of overclocking, you can select slightly higher voltages.

Finished with the Soyo, I will try the same technique on a different motherboard, an ACorp with VIA VPX chipset.

This weekend, I might be able to repair an ASUS A7V400-MX REV: 1.03, I had suspended the repairs, because with the replacement diode, present on a scrap MB, I had difficulty removing it, so I decided to order some new ones, and today they arrived, in total there are 50 pieces, surely it is a spare that will last forever (at least for me), they are BAT54C L43 SOT-23, these acronyms will not be useful to many, but I write it for those who maybe understand it, on the MB was missing D5, the marking on the diode was L43, surely there are other acronyms that Correspond to L43 (equivalents), I will have to take one and weld it to the MB, the operation fortunately should be possible, since there are no other components nearby.

Beyond that there should be some tracks to fix, but those I should look at them well under the microscope, if I find them scratched but not interrupted, I will have to add the heatsink to the NB, and then the board would be ready for the first boot, and since the chipset integrates VGA LAN and Audio, it will be enough to add a RAM and a CPU with heatsink and fan, to make it work, at least hopefully.

Today if I find the time ⏱ necessary, I could finish the work on the ASUS A7V400-MX REV: 1.03, we have to weld the D5, an SMD diode signed L43, the size is quite tiny, they don't make the work easy, however I think I can do it, because fortunately the area around it is quite free.

The next job, if this card does not require further work, I could do it on some Socket 7, I have already tried in the past to fix them, but today my knowledge of the problems and related solutions, are better, and so I could be able to make some work, among these I have a P5F76, i430VX chipset, which stopped working after a change of RAM, it is not a drama that no longer works, but it could be a nice challenge to repair it, originally it had an AMD K5-100, which I no longer used, like all the other K5s in my collection, the most Fast is a K5-166, which goes to about 116 MHz, decidedly efficient with integers, very deficient in decimals, even if maybe it is a little better than the Cyrix equivalents, if I could make this P5F76 work, I could use the K5-166, even if the jumpers allow the use of dual voltage CPUs, from 2.4V up to 3.2V, so I would seriously think about a K6-2 266 2.2V@2.4V, but I don't know how the linear regulator will react, theoretically when the frequency goes up the heat produced will rise, so 266 would be much hotter than 200-233, but going down the voltage, From 2.9V-3.2V to 2.4V, I should have a cooler regulator (?), or the temperature difference is negligible (?).

The next three S.7 motherboards, on which I will work, are these:

A PCChips M560 chipset TXPRO (aka ALI Aladin IV+), I kept away from it because I always preferred those with Intel chipset, however it could be interesting to make some bench, even if I doubt that it can get excellent results, for the rest it is quite similar to the i430TX, according to the manual the maximum is 266 MHz, but I think you can go further, and the highest FSB is 83 MHz, while the minimum VCORE seems to be 2.5V, I don't know if you can go further down, trying other combinations of jumpers.

Then there is an ASUS VX97, I have a fair number of i430VX cards, but none of the ASUS, a brand that I have been fond of since the first 386DX that I bought, here apart from the absence of SDRAM slots, it is possible that the K6 is also supported, in the manual there is the K5 as an AMD CPU, maybe 🤔 was not yet ⚓️ out, the voltage 😬 minimum, it is also here 2.5V, but the same is true 🎤 done before, maybe 🤔 changing the position of the jumpers, it is possible to find secret settings of the VCORE, according to the manual, the maximum FSB is 66, it may be possible that There is the possibility of other frequencies, including 75 MHz or even 83 MHz, depends on the frequency generator installed.

The latest is a Chaintech 5TDM2 M101, which has automatic recognition of the CPU, any modification must be made by BIOS, and it should be quite similar to the system used by QDI, looking at the manual, there is no minimum VCORE specified, there is only that of the CPU, which goes from 2.8V to 2.9V and 3.2V, while for the FSB it is written, that it is possible to select 68 75 and 83 MHz, I will not rule out that with an updated BIOS, it is possible to install a K6-2 or a K6-3, or even the equivalent plus versions.

I just finished the work on the ASUS A7V400-MX, I re-solded the SMD diode (D5) initialed L43, the operation itself is very simple, in reality it is not like that, the tiny dimensions complicate the work, and the final result, even if aesthetically is not what I wanted, it is not exactly identical to the factory one, but the important thing is to have soldered the 3 pins, and to have restored the functionality of that circuit, exactly I have no idea what the diode is connected to, but now the circuit is not interrupted, so if the scratched traces in the back have continuity, the board It could be ready for a boot test.

Before trying it, I have to check well with the microscope, some points seem to be scratched superficially, others instead it is not clear if in some places they can be so deep, to break the trace, as soon as I can I check them, if I don't find anything else to fix, I can start preparing the board for a first boot test.

PC@LIVE wrote on 2025-02-25, 16:53:

The next three S.7 motherboards, on which I will work, are these: […]
Show full quote

The next three S.7 motherboards, on which I will work, are these:

A PCChips M560 chipset TXPRO (aka ALI Aladin IV+), I kept away from it because I always preferred those with Intel chipset, however it could be interesting to make some bench, even if I doubt that it can get excellent results, for the rest it is quite similar to the i430TX, according to the manual the maximum is 266 MHz, but I think you can go further, and the highest FSB is 83 MHz, while the minimum VCORE seems to be 2.5V, I don't know if you can go further down, trying other combinations of jumpers.

Then there is an ASUS VX97, I have a fair number of i430VX cards, but none of the ASUS, a brand that I have been fond of since the first 386DX that I bought, here apart from the absence of SDRAM slots, it is possible that the K6 is also supported, in the manual there is the K5 as an AMD CPU, maybe 🤔 was not yet ⚓️ out, the voltage 😬 minimum, it is also here 2.5V, but the same is true 🎤 done before, maybe 🤔 changing the position of the jumpers, it is possible to find secret settings of the VCORE, according to the manual, the maximum FSB is 66, it may be possible that There is the possibility of other frequencies, including 75 MHz or even 83 MHz, depends on the frequency generator installed.

The latest is a Chaintech 5TDM2 M101, which has automatic recognition of the CPU, any modification must be made by BIOS, and it should be quite similar to the system used by QDI, looking at the manual, there is no minimum VCORE specified, there is only that of the CPU, which goes from 2.8V to 2.9V and 3.2V, while for the FSB it is written, that it is possible to select 68 75 and 83 MHz, I will not rule out that with an updated BIOS, it is possible to install a K6-2 or a K6-3, or even the equivalent plus versions.

Hi PC@LIVE,

A few comments on your Socket 7 boards.

Asus VX97:
There are indeed undocumented lower Vcore settings possible on this board.

VID0:3 jumper settings:
2.0V -> VID0: open, VID1: open, VID2: open
2.1V -> VID0: 1-2, VID1: open, VID2: open
2.2V -> VID0: open, VID1: 1-2, VID2: open
2.3V -> VID0: 1-2, VID1: 1-2, VID2: open
2.4V -> VID0: open, VID1: open, VID2: 1-2

When running a K6-2+/III+ on the VX97, use the 0204v patch J.2 BIOS. This BIOS also fixes the 32GB IDE HDD limit bug. This patched VX97 BIOS is on TRW and my website.

Chaintech 5TDM2:
Patching the BIOS for better CPU support for this jumperless board will be quite a challenge!

Having looked at jumperless socket 7 BIOSes from Abit, QDI, and Shuttle, I was surprised by the different implementations of the jumperless control on these boards. Clearly, the basic Award v4.5xPG core didn’t include jumperless control, so each motherboard manufacturer had to design their own solution.
QDI went very far with their SpeedEasy Technology, implementing the most automatic control of FSB, multiplier, and Vcore for each CPU model and stepping. Only the speed rating of the installed CPU had to be entered in the BIOS.

So I’m curious what I will find when analyzing the SeePU logic from Chaintech. 😉
I will sent you a PM when I have more to share about a possible patched BIOS for the 5TDM2.

Ciao, Jan

Chkcpu wrote on 2025-02-27, 21:25:

Hi PC@LIVE, […]
Show full quote

PC@LIVE wrote on 2025-02-25, 16:53:

The next three S.7 motherboards, on which I will work, are these: […]
Show full quote

The next three S.7 motherboards, on which I will work, are these:

A PCChips M560 chipset TXPRO (aka ALI Aladin IV+), I kept away from it because I always preferred those with Intel chipset, however it could be interesting to make some bench, even if I doubt that it can get excellent results, for the rest it is quite similar to the i430TX, according to the manual the maximum is 266 MHz, but I think you can go further, and the highest FSB is 83 MHz, while the minimum VCORE seems to be 2.5V, I don't know if you can go further down, trying other combinations of jumpers.

Then there is an ASUS VX97, I have a fair number of i430VX cards, but none of the ASUS, a brand that I have been fond of since the first 386DX that I bought, here apart from the absence of SDRAM slots, it is possible that the K6 is also supported, in the manual there is the K5 as an AMD CPU, maybe 🤔 was not yet ⚓️ out, the voltage 😬 minimum, it is also here 2.5V, but the same is true 🎤 done before, maybe 🤔 changing the position of the jumpers, it is possible to find secret settings of the VCORE, according to the manual, the maximum FSB is 66, it may be possible that There is the possibility of other frequencies, including 75 MHz or even 83 MHz, depends on the frequency generator installed.

The latest is a Chaintech 5TDM2 M101, which has automatic recognition of the CPU, any modification must be made by BIOS, and it should be quite similar to the system used by QDI, looking at the manual, there is no minimum VCORE specified, there is only that of the CPU, which goes from 2.8V to 2.9V and 3.2V, while for the FSB it is written, that it is possible to select 68 75 and 83 MHz, I will not rule out that with an updated BIOS, it is possible to install a K6-2 or a K6-3, or even the equivalent plus versions.

Hi PC@LIVE,

A few comments on your Socket 7 boards.

Asus VX97:
There are indeed undocumented lower Vcore settings possible on this board.

VID0:3 jumper settings:
2.0V -> VID0: open, VID1: open, VID2: open
2.1V -> VID0: 1-2, VID1: open, VID2: open
2.2V -> VID0: open, VID1: 1-2, VID2: open
2.3V -> VID0: 1-2, VID1: 1-2, VID2: open
2.4V -> VID0: open, VID1: open, VID2: 1-2

When running a K6-2+/III+ on the VX97, use the 0204v patch J.2 BIOS. This BIOS also fixes the 32GB IDE HDD limit bug. This patched VX97 BIOS is on TRW and my website.

Chaintech 5TDM2:
Patching the BIOS for better CPU support for this jumperless board will be quite a challenge!

Having looked at jumperless socket 7 BIOSes from Abit, QDI, and Shuttle, I was surprised by the different implementations of the jumperless control on these boards. Clearly, the basic Award v4.5xPG core didn’t include jumperless control, so each motherboard manufacturer had to design their own solution.
QDI went very far with their SpeedEasy Technology, implementing the most automatic control of FSB, multiplier, and Vcore for each CPU model and stepping. Only the speed rating of the installed CPU had to be entered in the BIOS.

So I’m curious what I will find when analyzing the SeePU logic from Chaintech. 😉
I will sent you a PM when I have more to share about a possible patched BIOS for the 5TDM2.

Ciao, Jan

Ciao Jan,
Thank you very much for your valuable advice, they can be very useful, especially those on the Asus VX97, they avoid an experimental work, always on that card I should do some tests, to find undocumented FSB settings (in the maximum 75 MHz manual), but I already know that it has a limit, which is that of the 72 PIN RAM, I don't think I have fast modules, I think I only have normal 8-16MB modules, the card despite the lack of SDRAM slots, I really like it, I thought I wouldn't find one, about twenty years ago I started the search, but the main obstacle was the excessive cost 💲, I have not yet decided which CPU to use, but from your suggestions, I would go with a K6-2-CXT, even if I would start with a Pentium 133, and then a P.MMX or a K6.
Currently I don't know what BIOS is present, but I will follow your advice, downloading and updating then to V.0204 J.2, first I would like to run some benches, mostly those of Phil's, maybe the results are equivalent.
For the Chaintech 5TDM2, the lack of jumpers can actually complicate things, exactly no idea 💡 of how the detection of the VCORE works, which I think is the most complicated part, I have a pair of S.7 jumperless cards, a QDI Titanium IB+ with your BIOS J.2, and a Siemens TX but in this motherboard, the automatic detection circuit, is absent (predisposed), I will make a mod. as soon as possible because it only works at 2.8V, and the K6-2 CXT needs less, the Phoenix BIOS sees it as a normal K6-300, actually from CPU-Z it looks correctly as K6-2-400 (6X66).

Unfortunately, one of the three S.7 cards, has a not easy damage to repair, at least speaking for me, it is a fairly impossible job, but with the right equipment and the replacement chip (assuming that it is not possible to redo the missing pins), it could be repaired.

Unfortunately it is the ASUS VX97, the SB chip has some missing pins, plus almost the whole row crooked, maybe it was an attempt to remove it (?), in any case I will keep the card aside, at the moment only for collection, in the future if I find the replacement SB, I could entrust it to someone who is able to replace it, hoping that there are no other problems beyond that.

Just out of curiosity, I will check the FSB frequencies of the ICS 9169CM, there could be some other frequency beyond those reported in the manual, apart from that, I see that the VCORE circuit is of the switching type, this was usually linear on many motherboards of the period.

I admit that the problem on the ASUS VX97, was a great disappointment, but for the replacement I thought it was necessary to find a scrap card, and take the chip, in reality this is not necessary, I just found out that they are for sale, and they are (I imagine) new, at the moment I do not order it, but I will do it later, the cost is not really cheap, but not exaggerated either.

As for the frequency generator, I saw a pdf, in which you can find other settings, among these there is the 83 MHz one, the utility is relative, but if necessary, you could do some bench, or some interesting overclock.

As suggested by Jan, the VCORE can select voltages for AMD CPUs, usually 2.2V or 2.4V, or even 2.0V or more, in addition its modified BIOS J.2, allows the use of large capacity disks, which is very useful if you recover discs of 40 GB or more.

I have the three S.7 boards between my legs, the one that interested me the most, a disappointment was detected, it's the ASUS VX97, the card is damaged in the SB chip, at the moment I'm putting it away, but in the future I'm thinking of getting a new one, and possibly with the help of someone better equipped than me, I would have it replaced.

Let's move on to a PCChips M560 (TXPRO = ALI A.4+), this one has a Dallas type chip, which must be fixed, currently it has been inserted in reverse, for the rest everything seems to be in order.

The last card is a Chaintech 5TDM2, Intel i430TX chipset, it has no jumper for CPU selection (FSB VCORE Multi etc...), actually there are some jumpers, but they concern something else, it should have a BIOS like those of QDI or ABIT, able to automatically recognize the CPU and the voltage, you just have to enter the BIOS and choose the correct speed, because it usually starts initially with low frequencies, then if you want you could try if you can overclock.

Of the three S.7 motherboards, I have already put away the ASUS VX97, but in days I order the SB chip, I will see later who to entrust the card for replacement, theoretically it should not be too complicated to replace it, you could also do it yourself, but it is not convenient, you run the risk of irreparably damaging the board, personally I do not intend to buy other professional tools 🛠, because I would not use them, or I would use them occasionally.

The Chaintech 5TDM2 card, in my opinion is very interesting, I think I can try it as soon as possible, there are various advantages, plus as many disadvantages, let's start by saying that it is jumperless, not a big problem if it works, but you have to wonder which CPU does it work with?

According to the booklet (pdf online), it seems that it is not possible to use CPUs with VCORE less than 2.8V, it is not a problem if you want to use CPUs up to 233 MHz, but it becomes if I want to put a K6-2 400 (or higher), these CPUs usually work at 2.2V, power them at 2.8V or even 2.9V, for a very short period it can be done, but I wonder if it is possible from BIOS, to go below 2.8V?

But above all the automatic recognition system, how does it identify K6-2?

It is usually recognized as K6 (normal), but from previous experience on the QDI Titanium, without a mod to the BIOS, the maximum multi would be 5.5X, while the CXT versions could, select 2X to get the 6X.

For this I will do some tests, assuming that the card works, using a normal K6-233, which I will try to work, if from BIOS it is possible, with VCORE of about 2.0-2.5V, if it stores those lower voltages, and does not return them to the default ones (2.8-2.9V), I would be quite sure that a K6-2 or 3, can be a possible update.